Presentation delivered at MeCCSA 2013 by Dr. Vera Slavtcheva-Petkova, highlighting findings of a paper co-authored with Dr. Monica Bulger and Dr. Victoria Nash of the Oxford Internet Institute
Abstract: Moral panics about the Internet’s “harmful effects” have been ongoing in current years, recently exacerbated by a UK parliamentary inquiry into online child protection. Is there scientific evidence supporting these fears? Does the Internet harm children and especially their health? This paper will present the findings from a narrative review of more than 300 journal articles discussing the scale and scope of online harms that young people experience. We will first present the three main types of harms identified in the literature: health-related harms, sex-related harms and cyber-bullying. Then the paper will focus in more detail on the evidence about health-related harms incurred as a result of online risks. We not only identify the kinds of health harms discussed in the literature – mainly related to pro-eating disorder websites, self-injury websites and problematic Internet use – but we also investigate whether and how the researchers operationalize harm. We also scrutinize the research methods used in the studies, which tend to differ significantly among the disciplines studied.
Does the Internet harm children's health? A critical review of the evidence
1. Does
the
Internet
harm
children’s
health?
Dr
Vera
Slavtcheva-‐Petkova,
University
of
Chester
Dr
Monica
Bulger,
Dr
Victoria
Nash,
Oxford
Internet
Ins4tute,
University
of
Oxford
2. Contents
• Child
protec4on
and
online
harms:
Academic
and
policy
context
• Methods
• Scope,
scale
and
opera4onalisa4on
of
health-‐
related
harms
• Conclusions
3. Academic
context
• Growing
body
of
research
on
level
and
character
of
Internet
by
under
18s,
and
nature
of
risks
and
opportuni4es
experienced
(e.g.,
Livingstone
&
Haddon,
2009;
Schrock
&
Boyd,
2008)
• Research
suggests
that
risks
are
greatest
for
those
most
vulnerable
offline,
and
that
overall,
the
opportuni4es
of
Internet
use
outweigh
risks
(e.g.
Mitchell
et
al.,
2010,
Livingstone
et
al.,
2011)
• Real
lack
of
research
which
quan4fies
or
analyses
level
of
actual
harm
rather
than
poten4al
risk.
4. Policy
context
• Wide
array
of
policy
measures
across
Europe:
hotlines
for
repor4ng
child
abuse
images,
industry
codes
of
conduct
regula4ng
use
of
mobile
content
and
services,
and
increasing
provision
of
parental
controls
by
ISPs.
• Much
of
this
is
result
of
self
or
co-‐regula4on
• Possible
and
actual
conflicts
with
other
rights,
e.g.
freedom
of
expression,
legal
due
process.
• Policy
purportedly
informed
by
research
(e.g.
mul4-‐
stakeholder
UK
Council
for
Child
Internet
Safety),
but
suscep4ble
to
media
pressure
and
“moral
panics”
5. Perceived
risks
vs.
actual
harms
“While
new
discoveries
almost
always
have
both
benefits
and
disadvantages,
breathless
nega4ve
coverage
of
technology
frightens
parents,
prevents
teenagers
from
learning
responsible
use,
and
fuels
panics,
resul4ng
in
misguided
or
uncons4tu4onal
legisla4on”
(Marwick
2008).
6. Our
study
• Evidence
on
the
extent
of
harms
experienced
by
children
as
a
result
of
online
risks:
A
cri4cal
synthesis
of
research
• Funded
by
the
Oxford
University
Press's
Fell
Fund
• A
review
of
empirical
studies
of
harms
associated
with
young
people’s
(under
18s)
Internet
use,
published
in
English
between
1997-‐2012
+
interviews
with
key
stakeholders
7. Methods
• Keywords:
“harm
AND
Internet
AND
children”,
“harm
AND
Internet
AND
adolescents”,
“harm
AND
Internet
AND
minors”,
“harm
AND
Internet
AND
teens”
and
“harm
AND
Internet
AND
teenagers”
• Categories
for
inclusion:
1.Empirical
work
2.Published
in
peer-‐reviewed
journal
3.Main
focus
of
study
was
young
people
(aged
under
18)
4.Addressed
Internet
use
5.Addressed
incidents
of
harm
related
to
online
interac4ons
8. Methods
• Original
search
yielded
over
4,000
publica4ons
• Aher
a
first-‐pass
review,
corpus
narrowed
to
271
studies.
• 271
were
reviewed
but
only
148
ar4cles
were
fully
coded
aher
strict
applica4on
of
the
inclusion
criteria.
• Coding
framework
adapted
from
the
EU
Kids
Online
public
repository.
It
included
21
items,
including
details
of
method,
target
popula4on,
context
and
how
and
whether
harm
was
opera4onalised.
9. Three
categories
of
harms
• Three
main
categories
of
harms:
1.Health-‐related
harms:
63
ar4cles
1.Sex-‐related
harms:
49
ar4cles
1.Cyberbullying
–
36
ar4cles
12. Operationalization
of
harm
• Harm
is
opera4onalized
in
less
than
half
of
the
studies
–
44.5%
• Self-‐harm:
The
most
common
defini4on
(33.3%
of
all
studies
and
75%
of
those
that
opera4onalize
the
term)
• Examples:
• “Parasuicide”,
“self-‐mu4la4on”
or
“self-‐injury”
(Adams
et
al.,
2005,
p.
1293)
• “A
form
of
ac4vely
managed
self-‐destruc4ve
behavior
that
is
not
intended
to
be
lethal”,
which
“subsumes
an
extensive
range
of
behaviours”
such
as
self-‐mu4la4on,
self-‐injurious
behavior,
deliberate
self-‐harm
and
self-‐
wounding
(Murray
et
al.,
2008,
p.
29)
13. Other
deCinitions
of
harm
2. “Desensi4za4on
to
violence
in
real
life
and
impairment
in
the
process
and
outcome
of
moral
evalua4on”
3. Physical
harm
4. Emo4onal
harm
14. Scale
of
Pro-‐ED
harms
• Much
of
the
discussion
pertains
to
perceived
or
poten,al
harm
rather
than
evidence
of
actual
harm
• Pro-‐ea4ng
disorder
websites:
1.Even
“suppor4ve”
websites
contain
poten4ally
harmful
content
–
“pro-‐ED
websites
tend
to
be
perceived
as
suppor4ve
by
users,
but
instead
appear
to
exacerbate
or
main
users’
ea4ng
disorder
symptoms”
2.Three
“possible
risks”:
a.“Opera4on
under
the
guise
of
‘support’
b.Reinforcement
of
disordered
ea4ng
c. Preven4on
of
help-‐seeking
and
recovery”
(Rouleau
and
von
Ranson,
2011,
p.
525)
15. Scale
of
Pro-‐ED
harms
• 85%
of
the
pro-‐ea4ng
disorder
websites
contain
“thinspira?on”
material
–
images
of
very
thin
models
or
celebri4es
used
to
inspire
weight
loss,
70%
have
“?ps
and
tricks”
on
die?ng
and
fas?ng
or
purging,
laxa4ves
and
pills
(Borzekowski
et
al.,
2010)
• 96%
of
users
of
pro-‐ea4ng
disorder
websites
and
46.4%
of
users
of
pro-‐recovery
sites
report
“learning
new
weight
loss
or
purging
techniques”
(Wilson
et
al.,
2006,
p.
e1635)
• 19.2%
of
users
say
they
feel
the
pro-‐ea4ng
disorder
websites
are
harmful
to
them
because
they
encourage
the
disorders/compe44on
among
par4cipants
and
have
“nega4ve
impact
on
self-‐effect”
(Csipke
and
Horne,
2007,
p.
200)
16. Scale
of
Pro-‐ED
harms
• Talbot
(2010)
argues
that
viewing
pro-‐ea4ng
disorder
websites
is
linked
to
a
number
of
nega4ve
effects:
1.“Higher
levels
of
die4ng
and
exercise
2.Higher
levels
of
drive
for
thinness,
body
dissa4sfac4on
and
perfec4onism
3.Posi4ve
correla4on
between
viewing
pro-‐ED
websites,
disease
dura4on
and
hospitaliza4ons”
(p.
686)
BUT:
1.Viewing
pro-‐ED
websites
may
INCREASE
ea4ng
disorder
behaviour
but
MIGHT
NOT
CAUSE
it
2.More
research
is
needed
to
determine
if
these
websites
DO
HARM
and
if
so,
TO
WHOM
and
OF
WHAT
FORM
(Talbot,
2010,
p.
694)
17. Scale
of
self-‐harm
• A
“normalizing”
and
a
“pathologizing”
discourse
(Franzén
&
Goszén,
2011,
p.
279)
• Lisle
evidence
regarding
the
prevalence
of
self-‐
injurious
behavior
and
its
rela4onship
to
use
of
self-‐
harm
websites
or
forums
• 80%
of
the
users
of
self-‐harm
message
boards
are
found
to
be
between
14
and
20
years
old
(Whitlock,
Powers,
&
Eckenrode,
2006).
• The
typical
adolescent
self-‐injurer
is
female,
ohen
with
a
history
of
abuse
and
an
ea4ng
disorder,
most
commonly
cutng
her
arms
and
legs
and
hiding
it.
18. Scale:
pro-‐suicidal
sites
• Much
more
conclusive
in
the
claims
about
harms
• A
meta-‐study
(Durkee
et
al.,
2011)
concludes
that
“pro-‐suicide
websites
and
online
suicide
pacts”
are
“high-‐risk
factors
for
facilita4ng
suicidal
behaviours,
par4cularly
among
isolated
and
suscep4ble
individuals”
BUT
some
forums
provide
opportuni4es
for
people
to
meet
others
with
similar
experiences,
“wherein
their
thoughts
and
feeling
are
not
condemned
nor
lectured
about”
(p.
3944)
19. Scale
of
“Internet
addiction”
• 35%
of
people
with
Problema4c
Internet
experience
are
<18
• BUT
“virtually
all
of
the
Internet
problem
behaviors”
are
“extensions
of
problem
behaviors
that
pre-‐
existed
the
advent
of
the
Internet”
(Mitchell,
Becker-‐Blease,
&
Finkelhor,
2005,
p.
506)
• The
typical
problema4c
Internet
user
is
a
teenager
who
spends
most
of
his/her
4me
on
the
computer,
including
at
night.
He/she
is
socially
isolated
and
plays
games
12-‐14
hours
a
day
• A
counsellor:
“It’s
the
same
thing
as
with
drug
addicts”
(Acier
and
Kern,
2011,
p.
986)
20. Conclusions
• It
seems
indisputable
that
some
children
experience
a
variety
of
health
harms
as
a
result
of
using
the
Internet.
• Documented
examples
include
young
people
assisted
or
encouraged
in
their
suicide
asempts
aher
visi4ng
pro-‐suicide
forums
and
young
girls
encouraged
to
maintain
or
exacerbate
their
ea4ng
disorders
or
self-‐injurious
behavior.
• Rela4vely
low
numbers
–
specific
case
studies
based
on
interviews
with
health
professionals
and/or
medical
records
21. Conclusions
• Interes4ng
textual
analysis
studies
and
surveys,
sugges4ng
high
prevalence
of
poten4al
harms
and
risks
but
lisle
evidence
on
actual
harms,
especially
in
rela4on
to
low-‐risk
children
• For
example,
whether
and
how
are
healthy
children
affected
by
using
ea4ng
disorders
websites
–
are
they
“harmed”
by
the
pro-‐ED
websites?
• Limita4ons
of
studies:
reliance
on
surveys
–
they
tell
us
a
lot
about
risks
and
perceived
harms
but
lisle
about
actual
harms
• Limita4ons
of
our
study:
focus
strictly
on
harms
and
studies
that
use
the
term,
hence
poten4ally
relevant
studies
might
have
been
omised
22. Any
questions?
•Dr
Vera
Slavtcheva-‐Petkova,
v.petkova@chester.ac.uk
•Dr
Monica
Bulger,
monica.bulger@oii.ox.ac.uk
•Dr
Victoria
Nash,
victoria.nash@oii.ox.ac.uk